ABSTRACT
Enterobacter aerogenes is an agent of hospital-acquired infection that exhibits a remarkable resistance to β-lactam antibiotics during therapy. Five successive isolates of E. aerogenes infecting a patient and exhibiting a multiresistance phenotype to β-lactam antibiotics and fluoroquinolones were investigated. Among these clinical strains, four presented resistant phenotypes during successive imipenem and colistin treatments. The involved resistance mechanisms exhibited by the successive isolates were associated with alterations of the outer membrane that caused a porin decrease and lipopolysaccharide modifications.
Enterobacter aerogenes is a common agent of hospital-acquired infection. It exhibits a remarkable adaptive capability and easily acquires resistance to β-lactam antibiotics during therapy (2, 3, 6, 13). The existence of a prevalent resistant clone of E. aerogenes has been reported in France, Belgium, and Spain (7, 9, 14, 17, 28). In addition, antibiotic resistance of E. aerogenes is associated with a high crude fatality rate in infected patients in Belgian hospitals (14). In the last 5 years, it has been shown that clinical isolates of this species, which are naturally resistant to aminopenicillins, often express an extended-spectrum β-lactamase, TEM-24, which gives rise to resistance to β-lactam antibiotics (2, 3, 9, 10, 12, 23, 28). Moreover, E. aerogenes exhibits acquired resistance to other families of antimicrobial agents. Previous studies have reported that clinical strains exhibiting an efflux process are resistant to β-lactam antibiotics, quinolones, tetracycline, and chloramphenicol (26, 27). Drug efflux can be coincident with a drastic reduction in drug uptake due to the loss of porin in the outer membrane (17, 26).
Among factors resulting in decreased mortality and morbidity due to invasive infections by gram-negative bacilli, the early initiation of an appropriate antibiotic agent has been well demonstrated (15, 21). Due to extended resistance of gram-negative bacteria to almost all antibiotics, the use of colistin has been proposed as a treatment alternative (22, 24, 29, 32).
We report here the case of a patient who experienced an infection caused by E. aerogenes that became resistant to imipenem during treatment through a decrease in porin synthesis. We document the step-by-step in vivo emergence of imipenem- and polymyxin-resistant variants during sequential antibiotic therapy. This report reinforces the anxiety associated with the lack of research into new antibacterial agents (3, 31, 34).
The patient, a 60-year-old woman, underwent an operation on 23 November 2001 in the cardiothoracic surgery unit for myocardial revascularization using coronary artery bypass grafting. The patient suffered from insulin-dependent diabetes mellitus and presented anteroseptal myocardial infarction in September 2000. Evolution was uneventful when fever appeared on 6 December. Blood, urine, and liquid from thoracic drainage were sent to the microbiology laboratory. The patient received a combination of cotrimoxazole and ciprofloxacin. Two E. aerogenes isolates were obtained from urine and drainage samples. Unfortunately, only the drainage isolate was preserved (isolate A). The isolate was susceptible to cefepime, imipenem, colistin (polymyxin E), and gentamicin but exhibited significant levels of resistance to ceftazidime and norfloxacin (Fig. 1). The treatment was shifted to imipenem and gentamicin. Gentamicin was stopped after 10 days because of worsening renal function. After 25 days of treatment (including 15 days of imipenem monotherapy), two imipenem-resistant variants were successively isolated from a thoracic drainage sample (isolates B and C). These isolates were characterized by high resistance to imipenem and β-lactams. On 14 January, imipenem was replaced by colistin. There was an initial reduction in the amount of liquid from thoracic drainage, but it became more purulent later, in the beginning of February, and allowed isolation, on 9 February, of a novel E. aerogenes isolate that had recovered imipenem and cefepime susceptibility but exhibited significant colistin resistance (isolate D). Colistin was changed to gentamicin therapy alone. The condition of the patient had remained stable for 10 days when the patient died from multiple organ failure associated with septic shock. At that time, a last E. aerogenes isolate (isolate E) with an antibiotic resistance phenotype comparable to that of the previous isolate was obtained from her blood. All isolates conserved a phenotype of gentamicin susceptibility during the clinical course.
PFGE of XbaI-digested genomic DNAs from various E. aerogenes isolates. (A) Run A (190 V overnight with a pulse time ranging from 3 to 10 s). (B) Run B (180 V overnight with a pulse time ranging from 20 to 40 s).
MATERIALS AND METHODS
Bacterial strains, growth conditions, and antibiotic susceptibility tests.Five E. aerogenes isolates were obtained from one patient during the 11 weeks of evolution of the infection. The isolates were named A, B, C, D, and E (Fig. 1). The E. aerogenes ATCC 13048 strain was used as the standard strain. Bacteria were routinely grown in Luria-Bertani medium or in Mueller-Hinton broth.
Susceptibility to antibiotics was determined by a standard twofold broth microdilution method according to NCCLS guidelines. The results were expressed as MICs in micrograms per milliliter. The activities of selected antibiotics were evaluated in the presence of an efflux pump inhibitor, phenylalanine arginine β-naphthylamide (25), as described previously (17, 27).
Epidemiological typing. E. aerogenes isolate typing was carried out by pulsed-field gel electrophoresis (PFGE) with a CHEF DRII system (Bio-Rad SA, Ivry-sur-Seine, France) as previously described (23). Briefly, the E. aerogenes chromosomal DNA was digested with the restriction enzyme XbaI (40 U; 5 h at 37°C). Electrophoresis was performed at 190 V overnight with a pulse time ranging from 3 to 10 s for the first run and at 180 V overnight with a pulse time ranging from 20 to 40 s during the second run. Ten clinical isolates (EA3, EA31, EA5, EA27, EA17, EA20, EA44, EA54, EA117, and EA119) of E. aerogenes were randomly chosen from our collection (7, 10, 17, 26) and used for PFGE profile comparison, in addition to the ATCC 13048 strain.
SDS-polyacrylamide gel electrophoresis and immunoblotting.Bacteria in the exponential growth phase in Luria-Bertani broth or in Mueller-Hinton broth were pelleted and solubilized at 96°C as described previously (26). Samples (equivalent to 0.02 optical density units at 600 nm) were loaded on sodium dodecyl sulfate (SDS)-polyacrylamide gels (11% polyacrylamide, 0.1% SDS) and subjected to electrophoresis. After electrotransfer to nitrocellulose, the membranes were saturated by incubation overnight with Tris-buffered saline (17, 26) containing 10% skim milk powder at 4°C. Incubations with polyclonal antibodies directed against denatured porins; against the specific porin internal L3 loop, OmpA, or OmpX; and against the second antibody were carried out as previously reported (11, 17, 26, 27). The lipopolysaccharide (LPS) profile was determined after silver staining of the whole-cell preparation treated with protease K (20) and run on a SDS-polyacrylamide gel electrophoresis gel with Tricine buffer (4). To determine the protein and LPS profiles of the various isolates, similar amounts of samples were loaded, as previously described (17, 26).
RESULTS
Strain characterization.Five E. aerogenes isolates were obtained from the patient during the 11 weeks of evolution of the infection. The isolates, named A, B, C, D, and E, were compared by PFGE to 10 distinct E. aerogenes isolates chosen from our collection (Fig. 1). Because a single base mutation in the chromosomal DNA of an isolate is sufficient to introduce differences in three fragments of its restriction pattern, isolates with restriction patterns showing the same differences in one to three fragments were considered to belong to the same genotype (23, 33). The molecular typing analyses showed that all of the tested isolates exhibited closely related molecular typing patterns (Fig. 1), suggesting that these strains belong to the prevalent epidemiological type previously described in the hospitals of the Marseille area (6, 7, 13). A similar result was obtained by using random amplified polymorphic DNA analysis (data not shown).
In addition, we could detect about four distinct profiles exhibiting two or three differences in this PFGE analysis (Fig. 1). Based on these criteria, the ABCDE isolates segregated in a profile distinct from that of the other isolates. Moreover, we observed marked variation for a product migrating at ∼170 to 180 kbp in these isolates, and this band discriminated between the ADE profile and the BC profile. This result suggests that strains B, C, D, and E are emergent resistant variants that were selected for from the original strain during therapy.
Antibiotic susceptibilities of E. aerogenes isolates.The reference strain ATCC 13048 was used as a control. Antibiotic susceptibility testing was performed by the broth microdilution method using Mueller-Hinton broth according to NCCLS guidelines. The following antibiotics were studied: imipenem, cefepime, ceftazidime, colistin, polymyxin B, norfloxacin, ciprofloxacin, cotrimoxazole, and gentamicin. Two isolates, B and C, showed identical profiles, with significant levels of resistance to imipenem and cefepime and a conserved susceptibility to colistin and polymyxin B (Table 1). In contrast, isolates A, D, and E showed imipenem susceptibility. Isolates D and E exhibited remarkable resistance to colistin and polymyxin B molecules.
Antibiotic susceptibilities of E. aerogenes strains
To investigate a possible role of an efflux mechanism in the resistance profiles of the various isolates, we assayed phenylalanine arginine β-naphthylamide (PAβN), an efflux pump inhibitor previously used in clinical gram-negative multidrug-resistant strains (17, 19, 25). We have demonstrated its potential to increase antibiotic susceptibility (e.g., chloramphenicol and quinolone) in the E. aerogenes strains expressing active efflux (17, 27). The presence of PAβN in the incubation medium, at a concentration which did not affect the growth of the E. aerogenes strains tested, did not induce modifications of norfloxacin and colistin susceptibilities. This suggests the absence of a PAβN-sensitive active efflux in the tested clinical strains.
Analysis of outer membrane structure.As for porin expression, the pre-imipenem therapy isolate A showed positive immunodetection with antibodies directed against the major Omp36 porin (Fig. 2 and Table 2), indicating normal porin synthesis. In contrast, no immunorelated porin was detected in isolates B and C. The porin signal was recovered with the post-imipenem therapy strains D and E (Fig. 2 and Table 2). Similar results were obtained with antibodies directed against the internal L3 loop of porins, showing that the porin detected in isolates A, D, and E contained this functional channel domain (data not shown). Immunodetection of OmpA, a protein involved in the outer membrane architecture, was positive for all isolates. The two isolates B and C showed an overexpression of OmpX compared to ATCC 13048 and the A, D, and E strains. The isolates A, B, and C presented oligosaccharide profiles similar to that of strain ATCC 13048 (Fig. 2). The last two isolates, D and E, exhibited different LPS chain profiles, and the balance among the various LPS bands was modified, suggesting an alteration of the LPS structure in these clinical strains.
Porin immunodetection and LPS profile of the E. aerogenes strains. The antibiotic therapy received by the patient is indicated above the time line (CIP, ciprofloxacin; BAC, cotrimoxazole; IMI, imipenem; GEN, gentamicin; COL, colistin). The dates of isolation (month/day/year) are given under the line. The triangle and arrow indicate the positions of porin and LPS alterations, respectively. Only the relevant part of the gel is shown.
Evolution of E. aerogenes outer membrane components and LPS profile
DISCUSSION
During the last decade, E. aerogenes has emerged as an important hospital pathogen responsible for nosocomial respiratory tract infections (9, 28, 35). E. aerogenes strains isolated from hospitalized patients generally exhibit high resistance to broad-spectrum antibiotics (3, 6), and the emergence of extended-spectrum cephalosporin-resistant strains has been documented (6, 10). In France, an increasing number of clinical strains have produced plasmidic extended-spectrum β-lactamase, chromosomal cephalosporinase, and aminoglycoside acetyltransferase and remain susceptible only to imipenem and gentamicin (6, 10, 17). Malléa et al. have described clinical E. aerogenes strains presenting a complex resistance strategy associating β-lactamase production, porin deficiency, and active efflux (26). Consequently, antibiotic therapy of E. aerogenes infections in France is now frequently based on a combination of imipenem and gentamicin.
The emergence of E. aerogenes strains with decreased susceptibility to imipenem has been described, and Arpin et al. reported a frequency of 4.6% (2). In some cases, an alteration of outer membrane permeability has been reported in resistant clinical isolates (10, 26). A strong correlation was reported between the presence of the nonspecific major porin, Omp36, and the β-lactam susceptibility of E. aerogenes isolates: the absence of porin was always correlated with a β-lactam-resistant phenotype (6, 7, 10, 11, 17, 26). In contrast, the presence of a functional porin was generally associated with imipenem susceptibility (6, 10, 26). In the case we analyze, detection of the porin was negative after treatment by imipenem (isolates B and C) and became positive after treatment was stopped (isolates D and E). An overexpression of OmpX was observed in the B and C isolates. It is interesting that a high level of synthesis of E. aerogenes OmpX induces a noticeable decrease in porin expression (16) and that overexpression of OmpX has been reported in various resistant E. aerogenes clinical isolates exhibiting a lack of porin (17). These data suggest that imipenem treatment selects in vivo for strains that exhibit a downregulation of porin expression. The porin failure is subsequently removed when the imipenem treatment stops and OmpX expression conjointly decreases. The regulation of porin expression may involve the ompX regulation cascade as mentioned previously (1, 5, 16, 17, 30). Immunodetection of OmpA, which plays a key role in the outer membrane architecture (26), was positive for all isolates, suggesting that the alteration focused on porin expression in these strains.
Recently, the reuse of “old” drugs, such as colistin, has been proposed as an alternative or rescue therapy for patients infected by multidrug-resistant strains (22, 29, 32). In the case we report here, the last two isolates (D and E) had acquired resistance to colistin under therapy. They exhibited a modified balance of LPS patterns, suggesting an alteration of LPS after colistin was introduced. Interestingly, this modification did not alter outer membrane protein assembly, as shown by porin-OmpA detection and by recovery of imipenem and cefepime susceptibility. A modification in LPS was previously demonstrated to be sufficient to confer significant polymyxin resistance on LPS mutants of E. coli and Salmonella enterica serovar Typhimurium (8, 18). In addition, the D and E isolates exhibited PFGE profiles similar to that of the A isolate. This may reflect selection of the A variant in response to drug use and the evolution to B-C and D-E. The A isolate may have persisted throughout at some sequestered site and may have been the direct progenitor of both B-C and D-E; this notion is supported by the molecular typing analyses, which indicate the closely related profiles of the variants.
Our results illustrate the flexibility of the responses of some bacterial pathogens to antibiotics. Colistin or imipenem alone was not efficient to avoid emergence of resistant E. aerogenes variants. As a consequence, with the recovery of imipenem susceptibility observed in the final isolates, we suggest the use of a combination therapy, e.g., imipenem-colistin, for severe E. aerogenes infection to circumvent the emergence of resistance mechanisms.
ACKNOWLEDGMENTS
We thank J. Chevalier, A. Davin-Regli, and C. Bornet for valuable discussions and B. Giumelli for expert assistance.
This work was supported by the Université de la Méditerranée, CNRS, and the Contrat de Recherche Clinique (Assistance Publique-Hôpitaux de Marseille).
FOOTNOTES
- Received 5 April 2004.
- Returned for modification 26 July 2004.
- Accepted 28 November 2004.
- Copyright © 2005 American Society for Microbiology
